隨著近年資訊技術的蓬勃發展，研發技術不斷的進步與創新，因此企業越來越重視商標建立來保護自身權利，以及利用商標來凸顯企業的品牌價值。但在全球商標註冊數量快速成長的過程中，常會有構成商標侵權的事件發生，而且大部分的原商標擁有者很難立即察覺而間接造成其商業利益的損害。此外，企業也時常為了防止自身陷於商標侵權的疑慮，而付出相當大的心力。而以往企業在衡量自身尚未註冊之商標的可行性時，常需要人力去查找現有已註冊之商標是否與企業所提出的商標過於類似，此工作往往耗費大量時間與人力，而且也可能會因遺漏而造成衡量效果不佳。因此本研究將基於商標法中的混淆之虞審查基準來進行商標相似性判定方法流程及系統建構。利用人工智慧下深度學習之自然語言處理與深度類神經網路進行商標相似性分析及衡量，其中包含卷積神經網路、Siamese 神經網路、向量空間模型等演算法的精進，針對大量商標資料之拼字、發音和圖像等多元特徵進行提取，並能提取出相似度高的商標資訊來給使用者作為參考。本研究希望藉由以上方法更精準的提取相似之商標資訊，也可讓欲申請商標的使用者提前了解當前市場相似商標設計的狀況，以避免商標過於相似導致侵權或是註冊失敗的情形發生。亦可讓既有商標所有權人探勘其商標是否遭他人非法或混淆使用之虞。

The rapid development of short product life cycle products, along with the continuous advancement and innovation of IT and R&D technology, require companies to be more proactive to protect their intellectual property rights, trademarks are intangible assets that greatly boast the financial worth and customer loyalty of companies. Given the rapid growth in the number of global trademark registrations, trademark infringement is increasing, and original trademark owners have difficulties protecting their IP which damages their commercial interests. This research constructs a trademark similarity assessment process based on the criteria for likelihood of confusion in trademark law, including spelling similarity, pronunciation similarity and image similarity, and uses natural language processing and deep learning methods to measure the similarity. Algorithms such as convolutional neural networks, siamese neural networks and vector space models are used to analyze the large numbers of trademarks. Finally, trademark information with high similarity is extracted and recommended to the user. In addition, the research verifies and evaluates the system using real cases of trademark infringement. This novel assessment method extracts similar trademark information accurately, and warns users who want to apply for similar trademarks know the conditions of potential infringement that will lead to registration failure.

中文摘要 I
Abstract II
致謝 III
List of Figures VI
List of Tables VII
1. Introduction 1
1.1 Research background and motivation 1
1.2 Research scope and purpose 3
1.3 Research framework and procedure 4
2. Literature review 6
2.1 Trademark infringement 6
2.2 Artificial intelligence 10
2.3 Spelling-based similarity techniques 13
2.4 Pronunciation-based similarity techniques 16
2.5 Image-based similarity techniques 18
3. Methodology 26
3.1 Method for spelling-based similarity 27
3.2 Method for pronunciation-based similarity 32
3.3 Method for image-based similarity 35
3.4 Comprehensive trademark infringement judgment 42
3.5 Evaluation 43
4. Case study 44
4.1 Key attributes of trademark infringement case 44
4.2 Key criteria for review in trademark infringement cases 45
4.3 Trademark infringement cases collection 46
4.4 Trademark infringement cases for spelling similarity 51
4.5 Trademark infringement cases for pronunciation similarity 55
4.6 Trademark infringement cases for image similarity 59
4.7 Text mining and analysis of trademark infringement cases 63
5. Conclusions 68
References 71
Appendix A - Examples of infringement cases for spelling similarity 79
Appendix B - Examples of infringement cases for pronunciation similarity 85
Appendix C - Examples of infringement cases for image similarity 89

[1] Anuar, F. M., Setchi, R., & Lai, Y. K. (2013). Trademark image retrieval using an integrated shape descriptor. Expert Systems with Applications, 40(1), 105-121.
[2] Bell, S., & Bala, K. (2015). Learning visual similarity for product design with convolutional neural networks. ACM Transactions on Graphics, 34(4). Retrieved from https://www.cs.cornell.edu/~kb/publications/SIG15ProductNet.pdf.
[3] Gupta, N., & Athavale, V. A. (2011). Comparative study of different low level feature extraction techniques for content based image retrieval. International Journal of Computer Technology and Electronics Engineering (IJCTEE), 1(1), 39-42.
[4] Lin, K., Yang, H., Hsiao, J., & Chen, C. (2015). Deep learning of binary hash codes for fast image retrieval. 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Boston, MA, 7-12 June, USA, 27-35.
[5] Liu, H., Tian, Y., Wang, Y., Pang, L., & Huang, T. (2016). Deep relative distance learning: tell the difference between similar vehicles. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, Nevada, USA, 27-30 June, 2167-2175.
[6] Roy, K., & Mukherjee, J. (2013). Image similarity measure using color histogram, color coherence vector, and sobel method. International Journal of Science and Research (IJSR), 2(1), 538-543.
[7] Tursun, O., & Kalkan, S. (2015). METU dataset: A big dataset for benchmarking trademark retrieval. 2015 14th IAPR International Conference on Machine Vision Applications (MVA), Tokyo, Japan, 18-22 May, 514-517.
[8] Wang, J., Song, Y., Leung, T., Rosenberg, C., Wang, J., Philbin, J., . . . Wu, Y. (2014). Learning fine-grained image similarity with deep ranking. 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, Ohio, USA, 23-28 June, 1386-1393.
[9] You, F., & Zhang, C. (2010). The technique of color and shape-based multi-feature combination of trademark image retrieval. 2010 International Conference on Multimedia Technology, Ningbo, China, 29-31 October, 1-5.
[10] Zagoruyko, S., & Komodakis, N. (2015). Learning to compare image patches via convolutional neural networks. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7-12 June, 4353-4361.
[11] Ronkainen, A. (2013). Intelligent trademark analysis: experiments in large-scale evaluation of real-world legal AI. 2013 14th Conference on Artificial Intelligence and Law, Rome, Italy, 10-14 June, 227-231.
[12] Scott, C. D. (2013). Trademark strategy in the internet age: Customer hijacking and the doctrine of initial interest confusion. Journal of Retailing, 89(2), 176-189.
[13] Fontan, L., Ferrané, I., Farinas, J., Pinquier, J., & Aumont, X. (2016). Using phonologically weighted Levenshtein distances for the prediction of microscopic intelligibility. In Annual conference Interspeech (INTERSPEECH 2016), San Francisco, CA, United States, 8-12 September, 650-654.
[14] Sanghavi, S. S., & Karwande, V. S. (2018). Trademark Approval System Using Semantic Similarity Computation. International Journal of Advance Scientific Research and Engineering Trends, 3(2), 1-4.
[15] Cheng, N., Yu, Z., & Wang, K. (2017). String similarity computing based on position and cosine. 2017 IEEE 7th International Conference on Electronics Information and Emergency Communication (ICEIEC), Macau, China, 21-23 July, 256-261.
[16] Parmar, V. P., & Kumbharana, C. K. (2014). Study existing various phonetic algorithms and designing and development of a working model for the new developed algorithm and comparison by implementing it with existing algorithm(s). International Journal of Computer Applications, 98(19), 45-49.
[17] Melekhov, I., Kannala, J., & Rahtu, E. (2016). Siamese network features for image matching. 2016 IEEE 23rd International Conference on Pattern Recognition (ICPR), Cancun, Mexico, 4-8 December, 378-383.
[18] Gomaa, W. H., & Fahmy, A. A. (2013). A survey of text similarity approaches. International Journal of Computer Applications, 68(13), 13-18.
[19] Appalaraju, S., & Chaoji, V. (2017). Image similarity using deep CNN and curriculum learning. arXiv preprint. Retrieved from https://arxiv.org/abs/1709.08761.
[20] Zhang, S., Hu, Y., & Bian, G. (2017). Research on string similarity algorithm based on Levenshtein Distance. 2017 IEEE 2nd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), Chongqing, China, 25-26 March, 2247-2251.
[21] Wang, H., Wen, Y., Xian, Y., & Li, R. (2016). Chinese sentence similarity calculation based on multi-features fusion. Journal of Residuals Science & Technology, 13(7). 213.1-213.6.
[22] Kostakis, O., & Papapetrou, P. (2015). Finding the longest common sub-pattern in sequences of temporal intervals. Data mining and knowledge discovery, 29(5), 1178-1210.
[23] Xiao, Y., Cui, R., & Huai, L. (2015). New method for calculating string similarity fusing location information. Application Research of Computers, 32(11), 3287-3290.
[24] Anuar, F. M., Setchi, R., & Lai, Y. K. (2016). Semantic retrieval of trademarks based on conceptual similarity. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 46(2), 220-233.
[25] Anuar, F. M., Setchi, R., & Lai, Y. K. (2013). A conceptual model of trademark retrieval based on conceptual similarity. Procedia Computer Science, 22, 450-459.
[26] Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint. Retrieved from https://arxiv.org/abs/1409.1556.
[27] Krizhevsky, A., & Hinton, G. (2009). Learning multiple layers of features from tiny image. Technical report, University of Toronto. Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.222.9220&rep=rep1&type=pdf.
[28] Koneru, K., Pulla, V. S. V., & Varol, C. (2016, July). Performance Evaluation of Phonetic Matching Algorithms on English Words and Street Names. 2016 5th International Conference on Data Management Technologies and Applications, Colmar, France, 24-26 July, 57-64.
[29] Philips, L. (2000). The double metaphone search algorithm. C/C++ users journal, 18(6), 38-43.
[30] Hood, D. (2004). Caverphone revisited. Technical Paper CTP150804. Retrieve from https://caversham.otago.ac.nz/files/working/ctp150804.pdf.
[31] Krizhevsky, A., Sutskever. I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, 1097-1105.
[32] Levenshtein, V. I. (1966). Binary codes capable of correcting deletions, insertions, and reversals. In Soviet physics doklady, 10(8), 707-710.
[33] WIPO. (2018). World Intellectual Property Indicators 2018. Retrieved from https://www.wipo.int/edocs/pubdocs/en/wipo_pub_941_2018.pdf
[34] TIPO. (2018). Agreement on trade-related aspects of intellectual property rights. Retrieved from https://www.tipo.gov.tw/ct.asp?xItem=207100&ctNode=6780&mp=1.
[35] Lanham. (2016). Lanham (trademark) act index. Retrieved from https://www.bitlaw.com/source/15usc/index.html.
[36] Eagle, P. (2016). News on patent, trade mark and design databases on the Internet. World Patent Information, 44, 10-11, 2016.
[37] Jiang, H., Han A. Q., WANG, M. J., Wang, Z., & Wu, Y. L. (2014). Solution algorithm of string similarity based on improved Levenshtein distance. Computer Engineering, 40(1), 222-227.
[38] Diao, X. C., Tan, M. C., & Cao, J. J. (2010). New method of character string similarity compute based on fusing multiple edit distances. Application Research of Computers, 27(12), 4523-4525.
[39] Barrón-Cedeno, A., Rosso, P., Agirre, E., & Labaka, G. (2010). Plagiarism detection across distant language pairs. 2010 23rd International Conference on Computational Linguistics, 37-45.
[40] Miller, G. A. (1995). WordNet: a lexical database for English. Communications of the ACM, 38(11), 39-41.
[41] Gupta, N., Athavale, V. A., & Khan, M. I. (2010). Graphical user interface of efficient image quality assessment using new similarity metrics. Digital Image Processing, 2(9), 342-347.
[42] Cao, W., Wang, X., Ming, Z., & Gao, J. (2018). A review on neural networks with random weights. Neurocomputing, 275, 278-287.
[43] Pan, Y. (2016). Heading toward artificial intelligence 2.0. Engineering, 2(4), 409-413.
[44] Liu, W., Wang, Z., Liu, X., Zeng, N., Liu, Y., & Alsaadi, F. E. (2017). A survey of deep neural network architectures and their applications. Neurocomputing, 234, 11-26.
[45] Schmidhuber, J. (2015). Deep learning in neural networks: An overview. Neural networks, 61, 85-117.
[46] Carrasquilla, J., & Melko, R. G. (2017). Machine learning phases of matter. Nature Physics, 13(5), 431-434.
[47] Jordan, M. I., & Mitchell, T. M. (2015). Machine learning: Trends, perspectives, and prospects. Science, 349(6245), 255-260.
[48] Yin, B. C., Wang, W. T., & Wang, L. C. (2015). Review of Deep Learning. Journal of Beijing University of Technology, 41(1), 48-59.
[49] Su, H., Zhu, X., & Gong, S. (2017). Deep learning logo detection with data expansion by synthesising context. 2017 IEEE Winter Conference on Applications of Computer Vision (WACV), 530-539.
[50] Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781. Retrieved from https://arxiv.org/abs/1301.3781.
[51] Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. Retrieved from https://arxiv.org/abs/1412.6980.
[52] Kondrak, G. (2003). Phonetic alignment and similarity. Computers and the Humanities, 37(3), 273-291.
[53] Learning Press, Inc. (2017). The big list of homophones. Retrieved from https://blog.allaboutlearningpress.com/homophones.
[54] World Trade Organization. (2017). WTO - A Handbook on the WTO TRIPS Agreement (Agreement on Trade-Related Aspects of Intellectual Property Rights), Retrieved from https://www.wto.org/english/res_e/publications_e/handbook_wtotripsag12_e.htm
[55] Raghavan, H., & Allan, J. (2004, May). Using soundex codes for indexing names in ASR documents. In Proceedings of the Workshop on Interdisciplinary Approaches to Speech Indexing and Retrieval at HLT-NAACL 2004, 22-27.
[56] Sayers, A. (2014). NYSIIS: Stata module to calculate nysiis codes from string variables. Retrieved from https://econpapers.repec.org/software/bocbocode/S457936.htm
[57] Koch, G., Zemel, R., & Salakhutdinov, R. (2015). Siamese neural networks for one-shot image recognition. In ICML Deep Learning Workshop (Vol. 2). Retrieved from https://www.cs.cmu.edu/~rsalakhu/papers/oneshot1.pdf
[58] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
[59] Zeiler, M. D., & Fergus, R. (2013). Visualizing and Understanding Convolutional Networks. arXiv preprint arXiv:1311.2901. Retrieved from https://arxiv.org/abs/1311.2901.
[60] Clarivate Analytics. (2019). CompuMark (Trademark research and protection). Retrieved from http://discover.clarivate.com/watch-trial-2019-tw?utm_campaign=EM_Trademark_Watch_Trial_CM_Taiwan_2019&utm_medium=email&utm_source=Eloqua.
[61] 劉柏宏 (Liu, B.-H)，2019，「商標侵權判例分析與判例推薦系統建立」，碩士論文(指導教授:張瑞芬)，國立清華大學工業工程與工程管理學系。